tests/phpunit/maintenance/DumpTestCase.php

   1 <?php
   2
   3 /**
   4  * Base TestCase for dumps
   5  */
   6 abstract class DumpTestCase extends MediaWikiLangTestCase {
   7
   8         /**
   9          * exception to be rethrown once in sound PHPUnit surrounding
  10          *
  11          * As the current MediaWikiTestCase::run is not robust enough to recover
  12          * from thrown exceptions directly, we cannot throw frow within
  13          * self::addDBData, although it would be appropriate. Hence, we catch the
  14          * exception and store it until we are in setUp and may finally rethrow
  15          * the exception without crashing the test suite.
  16          *
  17          * @var Exception|null
  18          */
  19         protected $exceptionFromAddDBData = null;
  20
  21         /**
  22          * Holds the XMLReader used for analyzing an XML dump
  23          *
  24          * @var XMLReader|null
  25          */
  26         protected $xml = null;
  27
  28         /**
  29          * Adds a revision to a page, while returning the resuting revision's id
  30          *
  31          * @param Page $page Page to add the revision to
  32          * @param string $text Revisions text
  33          * @param string $summary Revisions summary
  34          * @param string $model The model ID (defaults to wikitext)
  35          *
  36          * @throws MWException
  37          * @return array
  38          */
  39         protected function addRevision( Page $page, $text, $summary, $model = CONTENT_MODEL_WIKITEXT ) {
  40                 $status = $page->doEditContent(
  41                         ContentHandler::makeContent( $text, $page->getTitle(), $model ),
  42                         $summary
  43                 );
  44
  45                 if ( $status->isGood() ) {
  46                         $value = $status->getValue();
  47                         $revision = $value['revision'];
  48                         $revision_id = $revision->getId();
  49                         $text_id = $revision->getTextId();
  50
  51                         if ( ( $revision_id > 0 ) && ( $text_id > 0 ) ) {
  52                                 return [ $revision_id, $text_id ];
  53                         }
  54                 }
  55
  56                 throw new MWException( "Could not determine revision id (" . $status->getWikiText() . ")" );
  57         }
  58
  59         /**
  60          * gunzips the given file and stores the result in the original file name
  61          *
  62          * @param string $fname Filename to read the gzipped data from and stored
  63          *   the gunzipped data into
  64          */
  65         protected function gunzip( $fname ) {
  66                 $gzipped_contents = file_get_contents( $fname );
  67                 if ( $gzipped_contents === false ) {
  68                         $this->fail( "Could not get contents of $fname" );
  69                 }
  70
  71                 $contents = gzdecode( $gzipped_contents );
  72
  73                 $this->assertEquals(
  74                         strlen( $contents ),
  75                         file_put_contents( $fname, $contents ),
  76                         '# bytes written'
  77                 );
  78         }
  79
  80         /**
  81          * Default set up function.
  82          *
  83          * Clears $wgUser, and reports errors from addDBData to PHPUnit
  84          */
  85         protected function setUp() {
  86                 parent::setUp();
  87
  88                 // Check if any Exception is stored for rethrowing from addDBData
  89                 // @see self::exceptionFromAddDBData
  90                 if ( $this->exceptionFromAddDBData !== null ) {
  91                         throw $this->exceptionFromAddDBData;
  92                 }
  93
  94                 $this->setMwGlobals( 'wgUser', new User() );
  95         }
  96
  97         /**
  98          * Checks for test output consisting only of lines containing ETA announcements
  99          */
 100         function expectETAOutput() {
 101                 // Newer PHPUnits require assertion about the output using PHPUnit's own
 102                 // expectOutput[...] functions. However, the PHPUnit shipped prediactes
 103                 // do not allow to check /each/ line of the output using /readable/ REs.
 104                 // So we ...
 105
 106                 // 1. ... add a dummy output checking to make PHPUnit not complain
 107                 //    about unchecked test output
 108                 $this->expectOutputRegex( '//' );
 109
 110                 // 2. Do the real output checking on our own.
 111                 $lines = explode( "\n", $this->getActualOutput() );
 112                 $this->assertGreaterThan( 1, count( $lines ), "Minimal lines of produced output" );
 113                 $this->assertEquals( '', array_pop( $lines ), "Output ends in LF" );
 114                 $timestamp_re = "[0-9]{4}-[01][0-9]-[0-3][0-9] [0-2][0-9]:[0-5][0-9]:[0-6][0-9]";
 115                 foreach ( $lines as $line ) {
 116                         $this->assertRegExp(
 117                                 "/$timestamp_re: .* \(ID [0-9]+\) [0-9]* pages .*, [0-9]* revs .*, ETA/",
 118                                 $line
 119                         );
 120                 }
 121         }
 122
 123         /**
 124          * Step the current XML reader until node end of given name is found.
 125          *
 126          * @param string $name Name of the closing element to look for
 127          *   (e.g.: "mediawiki" when looking for </mediawiki>)
 128          *
 129          * @return bool True if the end node could be found. false otherwise.
 130          */
 131         protected function skipToNodeEnd( $name ) {
 132                 while ( $this->xml->read() ) {
 133                         if ( $this->xml->nodeType == XMLReader::END_ELEMENT &&
 134                                 $this->xml->name == $name
 135                         ) {
 136                                 return true;
 137                         }
 138                 }
 139
 140                 return false;
 141         }
 142
 143         /**
 144          * Step the current XML reader to the first element start after the node
 145          * end of a given name.
 146          *
 147          * @param string $name Name of the closing element to look for
 148          *   (e.g.: "mediawiki" when looking for </mediawiki>)
 149          *
 150          * @return bool True if new element after the closing of $name could be
 151          *   found. false otherwise.
 152          */
 153         protected function skipPastNodeEnd( $name ) {
 154                 $this->assertTrue( $this->skipToNodeEnd( $name ),
 155                         "Skipping to end of $name" );
 156                 while ( $this->xml->read() ) {
 157                         if ( $this->xml->nodeType == XMLReader::ELEMENT ) {
 158                                 return true;
 159                         }
 160                 }
 161
 162                 return false;
 163         }
 164
 165         /**
 166          * Opens an XML file to analyze and optionally skips past siteinfo.
 167          *
 168          * @param string $fname Name of file to analyze
 169          * @param bool $skip_siteinfo (optional) If true, step the xml reader
 170          *   to the first element after </siteinfo>
 171          */
 172         protected function assertDumpStart( $fname, $skip_siteinfo = true ) {
 173                 $this->xml = new XMLReader();
 174                 $this->assertTrue( $this->xml->open( $fname ),
 175                         "Opening temporary file $fname via XMLReader failed" );
 176                 if ( $skip_siteinfo ) {
 177                         $this->assertTrue( $this->skipPastNodeEnd( "siteinfo" ),
 178                                 "Skipping past end of siteinfo" );
 179                 }
 180         }
 181
 182         /**
 183          * Asserts that the xml reader is at the final closing tag of an xml file and
 184          * closes the reader.
 185          *
 186          * @param string $name (optional) the name of the final tag
 187          *   (e.g.: "mediawiki" for </mediawiki>)
 188          */
 189         protected function assertDumpEnd( $name = "mediawiki" ) {
 190                 $this->assertNodeEnd( $name, false );
 191                 if ( $this->xml->read() ) {
 192                         $this->skipWhitespace();
 193                 }
 194                 $this->assertEquals( $this->xml->nodeType, XMLReader::NONE,
 195                         "No proper entity left to parse" );
 196                 $this->xml->close();
 197         }
 198
 199         /**
 200          * Steps the xml reader over white space
 201          */
 202         protected function skipWhitespace() {
 203                 $cont = true;
 204                 while ( $cont && ( ( $this->xml->nodeType == XMLReader::WHITESPACE )
 205                         || ( $this->xml->nodeType == XMLReader::SIGNIFICANT_WHITESPACE ) ) ) {
 206                         $cont = $this->xml->read();
 207                 }
 208         }
 209
 210         /**
 211          * Asserts that the xml reader is at an element of given name, and optionally
 212          * skips past it.
 213          *
 214          * @param string $name The name of the element to check for
 215          *   (e.g.: "mediawiki" for <mediawiki>)
 216          * @param bool $skip (optional) if true, skip past the found element
 217          */
 218         protected function assertNodeStart( $name, $skip = true ) {
 219                 $this->assertEquals( $name, $this->xml->name, "Node name" );
 220                 $this->assertEquals( XMLReader::ELEMENT, $this->xml->nodeType, "Node type" );
 221                 if ( $skip ) {
 222                         $this->assertTrue( $this->xml->read(), "Skipping past start tag" );
 223                 }
 224         }
 225
 226         /**
 227          * Asserts that the xml reader is at an closing element of given name, and optionally
 228          * skips past it.
 229          *
 230          * @param string $name The name of the closing element to check for
 231          *   (e.g.: "mediawiki" for </mediawiki>)
 232          * @param bool $skip (optional) if true, skip past the found element
 233          */
 234         protected function assertNodeEnd( $name, $skip = true ) {
 235                 $this->assertEquals( $name, $this->xml->name, "Node name" );
 236                 $this->assertEquals( XMLReader::END_ELEMENT, $this->xml->nodeType, "Node type" );
 237                 if ( $skip ) {
 238                         $this->assertTrue( $this->xml->read(), "Skipping past end tag" );
 239                 }
 240         }
 241
 242         /**
 243          * Asserts that the xml reader is at an element of given tag that contains a given text,
 244          * and skips over the element.
 245          *
 246          * @param string $name The name of the element to check for
 247          *   (e.g.: "mediawiki" for <mediawiki>...</mediawiki>)
 248          * @param string|bool $text If string, check if it equals the elements text.
 249          *   If false, ignore the element's text
 250          * @param bool $skip_ws (optional) if true, skip past white spaces that trail the
 251          *   closing element.
 252          */
 253         protected function assertTextNode( $name, $text, $skip_ws = true ) {
 254                 $this->assertNodeStart( $name );
 255
 256                 if ( $text !== false ) {
 257                         $this->assertEquals( $text, $this->xml->value, "Text of node " . $name );
 258                 }
 259                 $this->assertTrue( $this->xml->read(), "Skipping past processed text of " . $name );
 260                 $this->assertNodeEnd( $name );
 261
 262                 if ( $skip_ws ) {
 263                         $this->skipWhitespace();
 264                 }
 265         }
 266
 267         /**
 268          * Asserts that the xml reader is at the start of a page element and skips over the first
 269          * tags, after checking them.
 270          *
 271          * Besides the opening page element, this function also checks for and skips over the
 272          * title, ns, and id tags. Hence after this function, the xml reader is at the first
 273          * revision of the current page.
 274          *
 275          * @param int $id Id of the page to assert
 276          * @param int $ns Number of namespage to assert
 277          * @param string $name Title of the current page
 278          */
 279         protected function assertPageStart( $id, $ns, $name ) {
 280
 281                 $this->assertNodeStart( "page" );
 282                 $this->skipWhitespace();
 283
 284                 $this->assertTextNode( "title", $name );
 285                 $this->assertTextNode( "ns", $ns );
 286                 $this->assertTextNode( "id", $id );
 287         }
 288
 289         /**
 290          * Asserts that the xml reader is at the page's closing element and skips to the next
 291          * element.
 292          */
 293         protected function assertPageEnd() {
 294                 $this->assertNodeEnd( "page" );
 295                 $this->skipWhitespace();
 296         }
 297
 298         /**
 299          * Asserts that the xml reader is at a revision and checks its representation before
 300          * skipping over it.
 301          *
 302          * @param int $id Id of the revision
 303          * @param string $summary Summary of the revision
 304          * @param int $text_id Id of the revision's text
 305          * @param int $text_bytes Number of bytes in the revision's text
 306          * @param string $text_sha1 The base36 SHA-1 of the revision's text
 307          * @param string|bool $text (optional) The revision's string, or false to check for a
 308          *            revision stub
 309          * @param int|bool $parentid (optional) id of the parent revision
 310          * @param string $model The expected content model id (default: CONTENT_MODEL_WIKITEXT)
 311          * @param string $format The expected format model id (default: CONTENT_FORMAT_WIKITEXT)
 312          */
 313         protected function assertRevision( $id, $summary, $text_id, $text_bytes,
 314                 $text_sha1, $text = false, $parentid = false,
 315                 $model = CONTENT_MODEL_WIKITEXT, $format = CONTENT_FORMAT_WIKITEXT
 316         ) {
 317                 $this->assertNodeStart( "revision" );
 318                 $this->skipWhitespace();
 319
 320                 $this->assertTextNode( "id", $id );
 321                 if ( $parentid !== false ) {
 322                         $this->assertTextNode( "parentid", $parentid );
 323                 }
 324                 $this->assertTextNode( "timestamp", false );
 325
 326                 $this->assertNodeStart( "contributor" );
 327                 $this->skipWhitespace();
 328                 $this->assertTextNode( "ip", false );
 329                 $this->assertNodeEnd( "contributor" );
 330                 $this->skipWhitespace();
 331
 332                 $this->assertTextNode( "comment", $summary );
 333                 $this->skipWhitespace();
 334
 335                 $this->assertTextNode( "model", $model );
 336                 $this->skipWhitespace();
 337
 338                 $this->assertTextNode( "format", $format );
 339                 $this->skipWhitespace();
 340
 341                 if ( $this->xml->name == "text" ) {
 342                         // note: <text> tag may occur here or at the very end.
 343                         $text_found = true;
 344                         $this->assertText( $id, $text_id, $text_bytes, $text );
 345                 } else {
 346                         $text_found = false;
 347                 }
 348
 349                 $this->assertTextNode( "sha1", $text_sha1 );
 350
 351                 if ( !$text_found ) {
 352                         $this->assertText( $id, $text_id, $text_bytes, $text );
 353                 }
 354
 355                 $this->assertNodeEnd( "revision" );
 356                 $this->skipWhitespace();
 357         }
 358
 359         protected function assertText( $id, $text_id, $text_bytes, $text ) {
 360                 $this->assertNodeStart( "text", false );
 361                 if ( $text_bytes !== false ) {
 362                         $this->assertEquals( $this->xml->getAttribute( "bytes" ), $text_bytes,
 363                                 "Attribute 'bytes' of revision " . $id );
 364                 }
 365
 366                 if ( $text === false ) {
 367                         // Testing for a stub
 368                         $this->assertEquals( $this->xml->getAttribute( "id" ), $text_id,
 369                                 "Text id of revision " . $id );
 370                         $this->assertFalse( $this->xml->hasValue, "Revision has text" );
 371                         $this->assertTrue( $this->xml->read(), "Skipping text start tag" );
 372                         if ( ( $this->xml->nodeType == XMLReader::END_ELEMENT )
 373                                 && ( $this->xml->name == "text" )
 374                         ) {
 375
 376                                 $this->xml->read();
 377                         }
 378                         $this->skipWhitespace();
 379                 } else {
 380                         // Testing for a real dump
 381                         $this->assertTrue( $this->xml->read(), "Skipping text start tag" );
 382                         $this->assertEquals( $text, $this->xml->value, "Text of revision " . $id );
 383                         $this->assertTrue( $this->xml->read(), "Skipping past text" );
 384                         $this->assertNodeEnd( "text" );
 385                         $this->skipWhitespace();
 386                 }
 387         }
 388 }